owner stringclasses 14
values | repo stringclasses 14
values | id int64 116k 4.61B | issue_number int32 1 181k | author stringlengths 1 39 | body stringlengths 1 262k | created_at timestamp[us]date 2000-06-06 02:40:44 2026-06-02 20:16:18 | updated_at timestamp[us]date 2000-06-06 02:40:44 2026-06-02 20:22:56 | reactions unknown | author_association stringclasses 5
values |
|---|---|---|---|---|---|---|---|---|---|
ClickHouse | ClickHouse | 820,328,155 | 23,113 | nvartolomei | #2592 | 2021-04-15T10:48:55 | 2021-04-15T10:48:55 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 819,366,979 | 22,706 | tavplubix | Please, take a look at previous comments again, some of them are not resolved | 2021-04-14T09:13:58 | 2021-04-14T09:13:58 | {} | MEMBER |
ClickHouse | ClickHouse | 819,429,831 | 23,044 | KochetovNicolai | Well, I have planned to read this pr before merge ... | 2021-04-14T10:58:59 | 2021-04-14T10:58:59 | {} | MEMBER |
ClickHouse | ClickHouse | 819,533,027 | 22,237 | alexey-milovidov | I've added the support. | 2021-04-14T13:49:17 | 2021-04-14T13:49:17 | {} | MEMBER |
ClickHouse | ClickHouse | 819,573,653 | 23,061 | DipalPrajapati | But i do not get the data as you can see. | 2021-04-14T14:43:53 | 2021-04-14T14:43:53 | {} | NONE |
ClickHouse | ClickHouse | 819,647,629 | 17,115 | alexey-milovidov | I've reproduced the issue with Google Data Studio.
Looks like it's easy to fix. | 2021-04-14T16:23:46 | 2021-04-14T16:23:46 | {} | MEMBER |
ClickHouse | ClickHouse | 819,814,367 | 23,085 | dankondr | @alexey-milovidov | 2021-04-14T20:35:23 | 2021-04-14T20:35:23 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 820,003,516 | 23,062 | marising | > Does it make sense to expand `CLUSTER PAUSE NODE` command, so such node would not participate in new Distributed queries?
>
> It would make easier to gracefully remove nodes without affecting any running query.
Distributed query is guaranteed by the load_balancing mechanism. If a replica node fails or the proce... | 2021-04-15T02:14:46 | 2021-04-15T02:14:46 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 820,011,681 | 21,337 | hexiaoting | @akuzm ready to review. | 2021-04-15T02:29:31 | 2021-04-15T02:29:31 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 820,227,943 | 22,299 | keen-wolf | > I tried to add a performance test, but selecting only one or several columns is extremely fast (and short queries should last at least 0.2s). So let's merge as is. BTW,
> <img alt="image" width="1698" src="https://user-images.githubusercontent.com/25705399/113869867-e0ba9400-97b9-11eb-874d-a19d3255cdf1.png">
> Seem... | 2021-04-15T08:25:54 | 2021-04-15T08:25:54 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 820,316,641 | 22,297 | amosbird | A great answer from @abyss7 in telegram
> The main course is to get rid of big context, split it into separate subclasses, like GlobalContext, SessionContext, QueryContext etc. and move global context to a singleton. As the result we will need no more to store context references everywhere.
> The new rules are:
>... | 2021-04-15T10:28:18 | 2021-04-15T10:28:18 | {} | COLLABORATOR |
ClickHouse | ClickHouse | 820,412,710 | 23,125 | akuzm | This is a bug, `limit` shouldn't be moved under the calculation of window functions. It's probably not very important, because the result of `limit` w/o `order by` is unspecified, and if you add an explicit `order by`, it works as expected:
```
WITH arrayJoin(['a', 'a', 'b', 'b']) AS field
SELECT
field,
co... | 2021-04-15T13:14:06 | 2021-04-15T13:14:06 | {
"+1": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 819,957,523 | 22,706 | songenjie | @tavplubix PTAL, the others failures are not related to the changes | 2021-04-15T01:41:21 | 2021-04-15T01:41:21 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 819,989,848 | 21,647 | den-crane | ```
......++++++++++........
.........+++++++++++++.. <----- the winner
...............+++++++++
^
|
searched
point
```
| 2021-04-15T02:08:18 | 2021-04-15T02:08:18 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 820,216,054 | 23,077 | alesapin | Ok now | 2021-04-15T08:07:59 | 2021-04-15T08:07:59 | {} | MEMBER |
ClickHouse | ClickHouse | 820,223,855 | 23,096 | Avogar | I will add a test | 2021-04-15T08:19:51 | 2021-04-15T08:19:51 | {} | MEMBER |
ClickHouse | ClickHouse | 820,240,673 | 23,069 | paeolo | @filimonov Got it, thanks a lot;) | 2021-04-15T08:44:05 | 2021-04-15T08:44:05 | {} | NONE |
ClickHouse | ClickHouse | 819,428,676 | 23,044 | alexey-milovidov | Differences in performance are unrelated. | 2021-04-14T10:56:44 | 2021-04-14T10:56:44 | {} | MEMBER |
ClickHouse | ClickHouse | 819,564,011 | 22,359 | alexey-milovidov | The patches are here:
https://github.com/zlib-ng/zlib-ng/compare/develop...ClickHouse-Extras:clickhouse
They don't look essential. | 2021-04-14T14:31:05 | 2021-04-14T14:31:05 | {} | MEMBER |
ClickHouse | ClickHouse | 819,246,302 | 22,760 | cfroystad | Thanks for the help!
As far as I can see (though I've been wrong before), the remaining failures and flakiness is unrelated to this change. | 2021-04-14T05:47:59 | 2021-04-14T05:47:59 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 819,562,415 | 22,359 | alexey-milovidov | Wow, it's working perfectly after library update.
But need to check for our patches and perf test. | 2021-04-14T14:29:02 | 2021-04-14T14:29:02 | {} | MEMBER |
ClickHouse | ClickHouse | 819,838,362 | 22,976 | alexey-milovidov | Now it's Ok:
```
wget https://clickhouse-builds.s3.yandex.net/22976/2ce494aa0fa9f9fd631c28340c443db7ad771eea/clickhouse_build_check/clang-11_relwithdebuginfo_none_bundled_unsplitted_disable_False_deb/clickhouse-server_21.5.1.6566_all.deb https://clickhouse-builds.s3.yandex.net/22976/2ce494aa0fa9f9fd631c28340c443db7ad... | 2021-04-14T21:16:36 | 2021-04-14T21:16:36 | {} | MEMBER |
ClickHouse | ClickHouse | 820,021,315 | 21,309 | amosbird | > @amosbird Why not create one distributed table and a bunch on VIEW's on top of it?
It's not easy to keep all these tables' metadata in sync. And iirc VIEW turns distributed table query into subqueries which will disable GLOBAL IN/JOIN and distributed IN/JOIN. | 2021-04-15T02:59:56 | 2021-04-15T02:59:56 | {} | COLLABORATOR |
ClickHouse | ClickHouse | 820,355,812 | 23,103 | tavplubix | Please take a look at `internal_replication` setting description: https://clickhouse.tech/docs/en/engines/table-engines/special/distributed/ | 2021-04-15T11:38:01 | 2021-04-15T11:38:01 | {} | MEMBER |
ClickHouse | ClickHouse | 819,430,332 | 23,056 | tavplubix | Full stack trace: https://gist.github.com/tavplubix/8b1edd51e03369a117e0451a9033be19
Segfault happened on `_exit` call [here](https://github.com/ClickHouse/ClickHouse/blob/495c6e03aa9437dac3cd7a44ab3923390bef9982/programs/server/Server.cpp#L1382). According to the man, it should terminate the calling process "immedi... | 2021-04-14T10:59:50 | 2021-04-14T10:59:50 | {} | MEMBER |
ClickHouse | ClickHouse | 819,494,460 | 23,061 | UnamedRus | It's expected Materialized View working as insert trigger actually.
https://github.com/ClickHouse/ClickHouse/issues/15295 | 2021-04-14T12:53:30 | 2021-04-14T12:53:30 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 819,497,865 | 23,036 | alexey-milovidov | https://github.com/ClickHouse/ClickHouse/issues/23056 | 2021-04-14T12:59:06 | 2021-04-14T12:59:06 | {} | MEMBER |
ClickHouse | ClickHouse | 819,566,192 | 22,359 | alexey-milovidov | `json_extract_simdjson` - will check this perf test manually. | 2021-04-14T14:33:51 | 2021-04-14T14:33:51 | {} | MEMBER |
ClickHouse | ClickHouse | 819,656,951 | 22,664 | abyss7 | Merged while it didn't get conflicts - will fix more tests in another PR. | 2021-04-14T16:35:49 | 2021-04-14T16:35:49 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 819,894,508 | 21,431 | alexey-milovidov | @danlark1 said that it just should be better. | 2021-04-14T22:38:14 | 2021-04-14T22:38:14 | {} | MEMBER |
ClickHouse | ClickHouse | 819,899,016 | 18,952 | alexey-milovidov | Recently @akuzm have found an issue with the usage of `ccache` in our CI.
After it was fixed most of the builds in CI becomes very fast (as they are started to be incremental).
For example, the binary build in master is only 3 minutes (the last row):
https://clickhouse-builds.s3.yandex.net/0/ca03670e60cfec19c84e1d... | 2021-04-14T22:50:15 | 2021-04-14T22:50:32 | {} | MEMBER |
ClickHouse | ClickHouse | 820,385,134 | 23,109 | den-crane | In the most cases it's correct.
In some cases it's not. For example you have a multi-tenant system and you have 10 000 tenants. You need to divide tenants because all queries have `where tenant = ... ` then you put tenant_id to first position in order_by. | 2021-04-15T12:30:42 | 2021-04-15T12:30:42 | {
"+1": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 820,433,749 | 22,299 | nikitamikhaylov | > @nikitamikhaylov you should copy changelog entry here, because the original PR was closed and won't show up in changelog. Also you'll have to mention the author manually in it.
@akuzm Done. | 2021-04-15T13:44:42 | 2021-04-15T13:44:42 | {} | MEMBER |
ClickHouse | ClickHouse | 819,361,840 | 23,058 | trollhe | i need multiple task to operation database,how i do? | 2021-04-14T09:06:32 | 2021-04-14T09:06:32 | {} | NONE |
ClickHouse | ClickHouse | 819,587,923 | 22,066 | alexey-milovidov | And the most flaky test is `test_materialize_mysql_database`, FYI @BohuTANG @zhang2014 | 2021-04-14T15:01:06 | 2021-04-14T15:01:06 | {} | MEMBER |
ClickHouse | ClickHouse | 819,627,183 | 21,309 | nikitamikhaylov | @amosbird Why not create one distributed table and a bunch on VIEW's on top of it? | 2021-04-14T15:54:32 | 2021-04-14T15:54:32 | {} | MEMBER |
ClickHouse | ClickHouse | 819,895,962 | 23,082 | alexey-milovidov | > Support FORMAT JSON for EXPLAIN PLAN query. Now, EXPLAIN with specified format other than JSON will throw exception (for old clients FORMAT is ignored). Omit FORMAT to get human-readable output.
But this logic is unorthogonal. | 2021-04-14T22:42:15 | 2021-04-14T22:42:15 | {} | MEMBER |
ClickHouse | ClickHouse | 819,904,434 | 22,802 | alexey-milovidov | Not relevant:
> PS. IMHO should be backported to >= 20.11. | 2021-04-14T23:03:22 | 2021-04-14T23:03:22 | {} | MEMBER |
ClickHouse | ClickHouse | 820,317,353 | 23,058 | trollhe | > @trollhe, this was fixed in version 21.4. (Fixed in #21839)
thanks | 2021-04-15T10:29:32 | 2021-04-15T10:29:32 | {} | NONE |
ClickHouse | ClickHouse | 819,208,525 | 21,318 | azat | >[AST fuzzer (TSan) — Lost connection to server. See the logs.
](https://clickhouse-test-reports.s3.yandex.net/21318/3afa94612a17d7a9501a37429c51a377d392ced5/fuzzer_tsan/report.html#fail1)
OOM/SIGKILL
>[Style Check — Style check failed.](https://clickhouse-test-reports.s3.yandex.net/21318/3afa94612a17d7a9501a374... | 2021-04-14T03:50:39 | 2021-04-14T03:51:08 | {} | MEMBER |
ClickHouse | ClickHouse | 819,215,672 | 13,889 | melin | @amosbird | 2021-04-14T04:16:40 | 2021-04-14T04:16:40 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 819,586,556 | 22,066 | alexey-milovidov | We have a dataset with all failed tests.
I've run a query:
```
SELECT test_name, count() AS c, max(check_start_time) AS last_time, argMin(commit_sha, check_start_time) AS first_seen, argMax(commit_sha, check_start_time) AS last_seen FROM `gh-data`.checks
WHERE
test_status = 'FAIL'
AND check_name LIKE 'I... | 2021-04-14T14:59:24 | 2021-04-14T14:59:56 | {} | MEMBER |
ClickHouse | ClickHouse | 817,467,257 | 22,222 | agusterodin | Has anyone able to get this docker image (altinity/clickhouse-server:21.4.1.6307-testing-arm) running on M1 Mac? Mine immediately crashes with runtime: failed to create new OS thread (have 2 already; errno=22. | 2021-04-12T04:20:59 | 2021-04-14T15:08:53 | {
"+1": 1
} | NONE |
ClickHouse | ClickHouse | 819,908,034 | 23,096 | alexey-milovidov | TODO add a test. | 2021-04-14T23:11:52 | 2021-04-14T23:11:52 | {} | MEMBER |
ClickHouse | ClickHouse | 820,010,880 | 23,061 | den-crane | > But i do not get the data as you can see.
Materialized View working as insert trigger. No insert, no MV triggered, no data.
`ENGINE = S3` does not do inserts.
| 2021-04-15T02:27:15 | 2021-04-15T02:27:15 | {
"+1": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 820,090,600 | 21,302 | keen-wolf | > This PR is merged in neighboring PR. All commits are saved.
great job! | 2021-04-15T04:43:37 | 2021-04-15T04:43:37 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 819,742,432 | 23,076 | aberrier | I found another issue related to this one.
It seems that even `toDecimal` will works or not 'correctly' depending if you query a nullable field or not.
```
SELECT toDecimal64('6.24994', 4),
toDecimal64('5.294345454499999', 4),
toDecimal64(toNullable('6.24994'), 4),
toDecimal64(toNullable('5... | 2021-04-14T18:33:42 | 2021-04-15T09:32:23 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 820,371,096 | 23,107 | alexey-milovidov | Please upgrade to version 21.4. | 2021-04-15T12:05:47 | 2021-04-15T12:05:47 | {} | MEMBER |
ClickHouse | ClickHouse | 819,174,663 | 23,037 | mjh568412187 | > The version 21.4 with Kerberos support was released yesterday. It is ready for production.
> You'd better to install it to your testing environment first, and only after testing for your own scenarios, deploy it on production.
>
> At Yandex we are using multiple different versions of ClickHouse (there are more th... | 2021-04-14T02:25:10 | 2021-04-14T02:25:10 | {} | NONE |
ClickHouse | ClickHouse | 819,470,341 | 7,649 | ccaapton | When I use conventional minhash for string similarity measure, like [datasketch](https://github.com/ekzhu/datasketch/) , I can use multiple permutation of the hash function to generate a fixed-length vector, and estimate the jaccard distance of two strings using hamming_distance/total_length.
But in clickhouse imple... | 2021-04-14T12:12:17 | 2021-04-14T12:12:17 | {
"+1": 1
} | NONE |
ClickHouse | ClickHouse | 819,695,536 | 21,858 | BoloniniD | I've moved YAML to XML translation function into a separate class and corrected the code in ConfigProcessor.cpp, the code looks better overall. | 2021-04-14T17:36:40 | 2021-04-14T17:36:40 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 819,901,667 | 18,952 | alexey-milovidov | So, this PR does two things:
1. Unity builds.
2. Faster linking.
Linking runs in several seconds if not IO bound. If it's IO bound it runs in the same time as needed to read and write several gigabytes on disk. So, it does not give much to optimize linking stage.
Unity builds make sense but:
\- the amount of c... | 2021-04-14T22:56:57 | 2021-04-14T22:57:27 | {} | MEMBER |
ClickHouse | ClickHouse | 819,947,300 | 17,455 | helifu | I think the "stat" command may have a hint
```
stat /clickhouse/tables/clarity/page_data_local_2/1-1/metadata
stat /clickhouse/tables/clarity/page_data_local_2/1-1/replicas/1/metadata
stat /clickhouse/tables/clarity/page_data_local_2/1-1/replicas/2/metadata
stat /clickhouse/tables/clarity/page_data_local_2/1-1/rep... | 2021-04-15T01:08:57 | 2021-04-15T01:10:57 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 820,152,437 | 23,069 | filimonov | > Does this feature interacts with the Kafka scenario described, ie. do we have block deduplication in the scenario described?
No, unfortunately not. Due to the possibility of consumer group rebalances it's not always possible to repeat exactly the same block.
There are some plans to address that issue. Currentl... | 2021-04-15T06:28:31 | 2021-04-15T06:28:31 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 820,353,102 | 22,359 | alexey-milovidov | Arcadia: `no such file or directory: '$(SOURCE_ROOT)/clickhouse/src/Dictionaries/LibraryDictionarySourceExternal.cpp'` | 2021-04-15T11:33:05 | 2021-04-15T11:33:05 | {} | MEMBER |
ClickHouse | ClickHouse | 815,658,074 | 22,841 | mastertheknife | About THP it does have some benefits, especially on x86_64 which doesn't handle high rate of TLB flushes as efficiently as aarch64 (from a personal experience with a VOIP server application).
I believe the issue here is that it tried to use 1GB pages. If it would have been 2MB pages it would probably be more accepta... | 2021-04-08T10:45:18 | 2021-04-14T07:34:57 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 819,439,303 | 20,058 | aberrier | > PS. But why do you have empty `<remote_url_allow_hosts></remote_url_allow_hosts>`?
Like @excitoon said :
> From documentation: `If this section is not present in configuration, all hosts are allowed`, real behavior was `If section contains no entries, all hosts are allowed`.
I believe in that time, we came to ... | 2021-04-14T11:15:18 | 2021-04-14T11:15:47 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 819,895,804 | 23,082 | alexey-milovidov | Submodule is unneeded. | 2021-04-14T22:41:48 | 2021-04-14T22:41:48 | {} | MEMBER |
ClickHouse | ClickHouse | 820,139,228 | 23,059 | millin | The same issue with 21.3.6.55 | 2021-04-15T06:02:24 | 2021-04-15T06:02:24 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 820,295,537 | 21,888 | Avogar | @rf I wrote some comments, please, take a look. | 2021-04-15T09:54:24 | 2021-04-15T09:54:24 | {} | MEMBER |
ClickHouse | ClickHouse | 820,354,198 | 23,088 | alexey-milovidov | There is overall 0.8% performance degradation. | 2021-04-15T11:35:04 | 2021-04-15T11:35:04 | {} | MEMBER |
ClickHouse | ClickHouse | 819,593,594 | 22,222 | agusterodin | If anyone else has this issue: try using lunalabsltd/clickhouse-server:20.6.7.4-arm as a stopgap. | 2021-04-14T15:08:45 | 2021-04-14T15:08:45 | {
"+1": 13
} | NONE |
ClickHouse | ClickHouse | 820,129,380 | 21,589 | dmalkr | It seems here is some problems with type inference. I've sent this query:
```sql
SELECT arrayFold( (x, acc) -> ([NULL], [NULL])
, [1, 2]
, (emptyArrayInt64(), emptyArrayInt64())
)
```
and server not failed but output
```
([78606152],[139752504390528])
```... | 2021-04-15T05:39:48 | 2021-04-15T05:39:48 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 820,209,964 | 23,108 | ikhudorozhkov | Need framing for special simbol in pass. | 2021-04-15T07:59:23 | 2021-04-15T07:59:23 | {} | NONE |
ClickHouse | ClickHouse | 820,429,176 | 23,138 | qoega | > Ok, but will it work when there is planned downtime in Sasovo?
It should work on any server. Also outside Yandex Intranet without ipv6. We need to check it before merge. | 2021-04-15T13:38:16 | 2021-04-15T13:38:16 | {} | MEMBER |
ClickHouse | ClickHouse | 819,814,144 | 23,047 | tavplubix | Stress test (undefined) - #23084
Yandex synchronization check - broken in master | 2021-04-14T20:34:57 | 2021-04-14T20:34:57 | {} | MEMBER |
ClickHouse | ClickHouse | 820,332,689 | 22,299 | akuzm | @nikitamikhaylov you should copy changelog entry here, because the original PR was closed and won't show up in changelog. Also you'll have to mention the author manually in it. | 2021-04-15T10:57:07 | 2021-04-15T10:57:07 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 820,198,544 | 23,108 | ikhudorozhkov | And on the end entered command --password DSDWSD!SFFWF | 2021-04-15T07:42:51 | 2021-04-15T13:19:11 | {} | NONE |
ClickHouse | ClickHouse | 819,375,801 | 22,706 | songenjie | > Please, take a look at previous comments again, some of them are not resolved
3ks, PTAL, I replied | 2021-04-14T09:27:54 | 2021-04-14T09:27:54 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 819,549,810 | 22,976 | alexey-milovidov | Unfortunately it keeps asking password even if the variable `DEBIAN_FRONTEND=noninteractive` is set.
(at least when terminal is available)
I can add a check for this variable in `Install.cpp`. | 2021-04-14T14:11:49 | 2021-04-14T14:11:49 | {
"confused": 1
} | MEMBER |
ClickHouse | ClickHouse | 819,587,689 | 22,066 | alexey-milovidov | From this result I see that the test `tests test_ttl_replicated` is being fixed. | 2021-04-14T15:00:44 | 2021-04-14T15:00:44 | {} | MEMBER |
ClickHouse | ClickHouse | 820,116,395 | 21,589 | dmalkr | When dealing with fuzzers output, I've found strange behavior of `insert()` function for tupes `(NULL, NULL)`.
Let's query this:
```sql
SELECT arrayFold( (x, acc) -> (NULL, NULL)
, [1]
, (emptyArrayInt64(), emptyArrayInt64())
)
```
Currently it is tend t... | 2021-04-15T05:23:57 | 2021-04-15T05:23:57 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 820,354,606 | 23,087 | alexey-milovidov | There is overall 0.9% performance degradation. | 2021-04-15T11:35:48 | 2021-04-15T11:35:48 | {} | MEMBER |
ClickHouse | ClickHouse | 820,395,243 | 21,319 | ka1bi4 | Hi @alexey-milovidov
This PR is ready for merge. | 2021-04-15T12:47:52 | 2021-04-15T12:47:52 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 819,368,508 | 23,047 | tavplubix | Stress test (thread) - segfault on `_exit` | 2021-04-14T09:16:23 | 2021-04-14T09:16:23 | {} | MEMBER |
ClickHouse | ClickHouse | 819,601,187 | 22,066 | alexey-milovidov | FYI `test_multiple_disks` is fixed in #16235. | 2021-04-14T15:18:50 | 2021-04-14T15:18:50 | {} | MEMBER |
ClickHouse | ClickHouse | 819,786,916 | 23,077 | alesapin | oops, no old logs for database replicated
```
var/lib/clickhouse2/coordination/
var/lib/clickhouse2/coordination/logs/
var/lib/clickhouse2/coordination/logs/changelog_820001_830000.bin
var/lib/clickhouse2/coordination/logs/changelog_810001_820000.bin
var/lib/clickhouse2/coordination/snapshots/
var/lib/clickhouse... | 2021-04-14T19:47:25 | 2021-04-14T19:47:25 | {} | MEMBER |
ClickHouse | ClickHouse | 820,357,459 | 23,038 | alesapin | Two days and I'm able to run bench via Jepsen on three nodes.
### Create requests with 32 bytes payload:
ZooKeeper + fsync
```
read requests 0, write requests 25802, Write RPS: 1430.95, Write MiB/s: 0.208793.
Write sampler:
0% 0.014457 sec.
10% 0.0163485 sec.
20% 0.01694... | 2021-04-15T11:41:06 | 2021-04-15T11:56:25 | {} | MEMBER |
ClickHouse | ClickHouse | 819,421,395 | 23,058 | kssenii | @trollhe, this was fixed in version 21.4. (Fixed in https://github.com/ClickHouse/ClickHouse/pull/21839) | 2021-04-14T10:43:23 | 2021-04-14T12:06:04 | {} | MEMBER |
ClickHouse | ClickHouse | 819,639,314 | 23,000 | alexey-milovidov | Need to fix Arcadia. | 2021-04-14T16:11:26 | 2021-04-14T16:11:26 | {} | MEMBER |
ClickHouse | ClickHouse | 819,808,407 | 23,041 | tavplubix | Integration tests, Performance - looks unrelated
SQLancer test - `Cannot parse NaN.: while converting 'N6\n?dZ' to Float64` - bug in SQLancer, it generates invalid queries
Stress test (thread) - #23056
Yandex synchronization check - broken in master | 2021-04-14T20:24:42 | 2021-04-14T20:24:42 | {} | MEMBER |
ClickHouse | ClickHouse | 819,898,358 | 18,952 | alexey-milovidov | @gonzalezjo This PR has a chance. | 2021-04-14T22:48:25 | 2021-04-14T22:48:25 | {} | MEMBER |
ClickHouse | ClickHouse | 819,903,871 | 22,861 | alexey-milovidov | I'm sure that integration tests failures are unrelated. | 2021-04-14T23:02:15 | 2021-04-14T23:02:15 | {} | MEMBER |
ClickHouse | ClickHouse | 820,167,159 | 22,594 | excitoon | @nikitamikhaylov the bug was in few places (not sure if I found every one), for example in `S3CredentialsProviderChain`:
```cpp
/// EC2MetadataService throttles by delaying the response so the service client should set a large read timeout.
/// EC2MetadataService delay is in order of ... | 2021-04-15T06:55:04 | 2021-04-15T06:57:34 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 820,326,267 | 22,762 | vdimir | I removed holding variant of function with Settings or without it and pass Settings to all aggregate function creators. | 2021-04-15T10:45:16 | 2021-04-15T10:45:16 | {} | MEMBER |
ClickHouse | ClickHouse | 820,352,641 | 22,359 | alexey-milovidov | After recompiling, the difference in simdjson vanished. | 2021-04-15T11:32:15 | 2021-04-15T11:32:15 | {} | MEMBER |
ClickHouse | ClickHouse | 820,358,347 | 23,000 | alexey-milovidov | Performance improvement is as expected. | 2021-04-15T11:42:48 | 2021-04-15T11:42:48 | {} | MEMBER |
ClickHouse | ClickHouse | 820,367,986 | 23,113 | alexey-milovidov | @nvartolomei This is very easy task, why don't do it? | 2021-04-15T12:00:16 | 2021-04-15T12:00:16 | {} | MEMBER |
ClickHouse | ClickHouse | 820,434,861 | 23,125 | wol-e | Thanks for the quick response and the simple work around 👍 | 2021-04-15T13:46:15 | 2021-04-15T13:46:15 | {} | NONE |
ClickHouse | ClickHouse | 819,627,780 | 21,304 | nikitamikhaylov | Maybe also add some test case from linked issue?
| 2021-04-14T15:55:23 | 2021-04-14T15:55:23 | {} | MEMBER |
ClickHouse | ClickHouse | 819,786,371 | 23,077 | alesapin | Ok, logs uploaded. Let's merge | 2021-04-14T19:46:25 | 2021-04-14T19:46:25 | {} | MEMBER |
ClickHouse | ClickHouse | 819,878,339 | 22,359 | alexey-milovidov | Two functions stopped to be inlined. But there is no difference on my machine.
Most likely it is due to ThinLTO. | 2021-04-14T22:04:37 | 2021-04-14T22:04:37 | {} | MEMBER |
ClickHouse | ClickHouse | 820,013,471 | 22,503 | hexiaoting | @CurtizJ ready to review.
The fail of Functional stateless tests flaky check seems that concurrent execution of my test case.
The rest error seems not related to my codes. | 2021-04-15T02:34:54 | 2021-04-15T02:34:54 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 820,187,025 | 23,082 | KochetovNicolai | > But this logic is unorthogonal.
I though that it is better to remove `json = 1` setting and use `FORMAT JSON` instead of.
As we discussed, it does not seem so useful to get json in json or json in xml. | 2021-04-15T07:27:50 | 2021-04-15T07:27:50 | {} | MEMBER |
ClickHouse | ClickHouse | 820,195,384 | 23,106 | zfno111 | sry wrong click | 2021-04-15T07:38:09 | 2021-04-15T07:38:09 | {} | NONE |
ClickHouse | ClickHouse | 820,286,598 | 22,745 | vdimir | Decided to merge revert #22753 | 2021-04-15T09:43:23 | 2021-04-15T09:43:23 | {} | MEMBER |
ClickHouse | ClickHouse | 812,394,791 | 20,200 | filimonov | expose following as counters:
- Parts
- InactiveParts
- ActiveParts
- SkippingIndexesGranulasScanned
- SkippingIndexesGranulasSkipped
| 2021-04-02T07:45:55 | 2021-04-02T07:51:33 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 812,482,532 | 22,475 | mastertheknife | Done fixing all of the above. | 2021-04-02T11:00:54 | 2021-04-02T11:00:54 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 812,527,650 | 14,896 | UnamedRus | No | 2021-04-02T13:24:16 | 2021-04-02T13:24:16 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 812,584,582 | 22,495 | haiertashu | > can you share verbose info curl -vvvv for both
I append the -vvv option to query local table like this
```
curl -o log -sSvvv 'http://uname:passwd@10.110.119.120:5566' -d 'select * from mydb.my_distribute_table_local'
```
The slow query by vip
```
run at Fri Apr 2 23:32:41 CST 2021
* About to connect() t... | 2021-04-02T15:43:17 | 2021-04-02T15:43:17 | {} | NONE |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.